When writing queries to >control, flags are parsed 1st, since they are
the only required field, and they require specific compositions. So
if the flags draw an error (on those specifics), then keyword errors
aren't reported. This can be mildly confusing/annoying, so explain it
instead.
cc: linux
[BONAIRE|HAWAII]_GB_ADDR_CONFIG_GOLDEN are only used by GFX7. So keep them
where they are needed.
Signed-off-by: Alexandre Demers
---
drivers/gpu/drm/amd/amdgpu/cikd.h | 3 ---
drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c | 3 +++
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drive
While I don't think it is being used anywhere, if it were used, it would
be wrong. We can base this assumption on MAX_POWER_MASK, where the shift is
by 16 bits.
Signed-off-by: Alexandre Demers
---
drivers/gpu/drm/radeon/sid.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/dr
[Public]
Reviewed-by: Kent Russell
> -Original Message-
> From: Cornwall, Jay
> Sent: Saturday, March 22, 2025 10:10 AM
> To: amd-gfx@lists.freedesktop.org
> Cc: Cornwall, Jay ; Russell, Kent
>
> Subject: [PATCH v2] drm/amdgpu: Increase KIQ invalidate_tlbs timeout
>
> KIQ invalidate_
The title's prefix should only indicate [PATCH]. It was part of the
series just sent a moment ago, but it was removed from it because it
concerned drm/radeon.
Alexandre
On Sat, Mar 22, 2025 at 2:39 PM Alexandre Demers
wrote:
>
> While I don't think it is being used anywhere, if it were used, it w
This should be titled patch 5/5. There was initially another one in
the lot about radeon. However, it was removed from the series and sent
alone.
Alexandre
On Sat, Mar 22, 2025 at 2:37 PM Alexandre Demers
wrote:
>
> Bring things on a single line and fix spacing.
>
> Signed-off-by: Alexandre Demer
cik_sdma_is_idle() does exactly what we need, so use it.
Signed-off-by: Alexandre Demers
---
drivers/gpu/drm/amd/amdgpu/cik_sdma.c | 6 +-
1 file changed, 1 insertion(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
index ede1a02
gmc_v7_0_is_idle() does exactly what we need, so use it.
Signed-off-by: Alexandre Demers
---
drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c | 9 +
1 file changed, 1 insertion(+), 8 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
index b601
Bring things on a single line and fix spacing.
Signed-off-by: Alexandre Demers
---
drivers/gpu/drm/amd/amdgpu/dce_v8_0.c | 10 +++---
drivers/gpu/drm/amd/amdgpu/vi.c | 8
2 files changed, 11 insertions(+), 7 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
b/
Tidy cik_sdma_hw_init() by returning directly cik_sdma_start()'s result.
Keep amdgpu_cik_gpu_check_soft_reset() early declaration with others.
Signed-off-by: Alexandre Demers
---
drivers/gpu/drm/amd/amdgpu/cik_sdma.c | 12 +++-
1 file changed, 3 insertions(+), 9 deletions(-)
diff --git
This series tidy up CIK code a bit.
It uses some functions already available to remove redundant code in GMC7
and CIK SDMA.
Some GOLDEN defines are moved into GFX7, the only place where they are used.
It fixes coding style in VI and DCE8.
Alexandre Demers (5):
drm/amdgpu: use gmc_v7_0_is_idle()
I should have written "v2" in the title
Alexandre
On Fri., Mar. 21, 2025, 21:47 Alexandre Demers, <
alexandre.f.dem...@gmail.com> wrote:
> The following series is intented to remove duplicated defines, shifts and
> masks or
> to classify them where they belong. si_enums.h has been used as a garb
Add new drm_dyndbg_user.c with a single call to
DYNDBG_CLASSMAP_USE(drm_debug_classes). This creates a _class_user
record (and a linkage dependency).
If a driver adds this object to its Makefile target, it gets the
record, which authorizes dyndbg to enable the module's class'd
pr_debugs, such as
KIQ invalidate_tlbs request has been seen to marginally exceed the
configured 100 ms timeout on systems under load.
All other KIQ requests in the driver use a 10 second timeout. Use a
similar timeout implementation on the invalidate_tlbs path.
v2: Poll once before msleep
Signed-off-by: Jay Cornw
On 3/21/2025 14:36, Russell, Kent wrote:
Do we want to actually do a fence_wait_polling before the loop begins like in
all of the other functions? I know 5ms isn't a big deal, but the other ones we
have are all after a single fence_wait_polling first.
Oh I see what you're saying now. This in
On 3/22/25 19:04, Ingo Molnar wrote:
>
> * Balbir Singh wrote:
>
>>> How frequently does this happen and what is the impact to users if
>>> this happens?
>>
>> It happens a 100% of the time when the BAR space lies beyond the
>> 10TiB region.
>
> And how frequently is the BAR space beyond the
Now that we are using the proper defines, cleanup useless old "substituted"
defines.
Signed-off-by: Alexandre Demers
---
drivers/gpu/drm/amd/amdgpu/sid.h | 1219 +-
1 file changed, 9 insertions(+), 1210 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/sid.h b/dr
It seems a copy-paste error: since we are working with
mmGRPH_SECONDARY_SURFACE_ADDRESS,
GRPH_SECONDARY_SURFACE_ADDRESS__GRPH_SECONDARY_SURFACE_ADDRESS_MASK
should be used.
Signed-off-by: Alexandre Demers
---
drivers/gpu/drm/amd/amdgpu/dce_v6_0.c | 2 +-
1 file changed, 1 insertion(+), 1 deleti
* Balbir Singh wrote:
> > How frequently does this happen and what is the impact to users if
> > this happens?
>
> It happens a 100% of the time when the BAR space lies beyond the
> 10TiB region.
And how frequently is the BAR space beyond the 10TiB region, on modern
systems?
Thanks,
Following the dyndbg-api-fix, replace DECLARE_DYNDBG_CLASSMAP with
DRM_CLASSMAP_USE. This refs the defined & exported classmap, rather
than re-declaring it redundantly, and error-prone-ly.
This resolves the appearance of "class:_UNKNOWN_" in the control file
for the driver's drm_dbg()s.
Fixes: f
20 matches
Mail list logo
